The 70-MPH speed limit: speed adaptation, spillover and surrogate measures of safety

(1)

2007

The 70-MPH speed limit: speed adaptation,

spillover and surrogate measures of safety

Victor Kenneth Lund

Iowa State University

Follow this and additional works at:https://lib.dr.iastate.edu/rtd

Part of theCivil Engineering Commons

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contactdigirep@iastate.edu.

Recommended Citation

Lund, Victor Kenneth, "The 70-MPH speed limit: speed adaptation, spillover and surrogate measures of safety" (2007).Retrospective Theses and Dissertations. 15054.

(2)

by

Victor Kenneth Lund

A thesis submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE

Major: Civil Engineering (Transportation Engineering) Program of Study Committee:

Reginald Souleyrette, Major Professor Thomas Maze

Thomas Stout Alicia Carriquiry

Iowa State University Ames, Iowa

2007

(3)

1446053 2007

UMI Microform Copyright

ProQuest Information and Learning Company 300 North Zeeb Road

P.O. Box 1346

Ann Arbor, MI 48106-1346

(4)

Disclaimer

This report was used in partial fulfillment of the requirements for the degree of Master of Science of the student researcher. The length of the after period used in this analysis is less than recommended by the 70-mph Steering committee. Additionally, at least 1,100 crashes were added to the Iowa Department of Transportation crash database after the download was completed for this research. As such, a thorough analysis could not be

performed due to limited time constraints on behalf of the student. The data analysis included herein is not intended to be representative of an exhaustive before and after study of the safety effects due to the increase of the rural interstate speed limit from 65-mph to 70-mph in Iowa and should therefore be considered preliminary. Future analysis completed on data that become available later may alter the results as reported herein.

(5)

List of Tables

Table 1. Summary of State Interstate Speed Limit Studies... 12

Table 2. Road Segment Assignment Scheme ... 29

Table 3. Summary of the Before and After Period Monthly Crash Frequency Means ... 68

Table 4. Summary of the Change in Crash Frequencies Adjusted by Traffic Volume ... 69

Table 5. Summary of All Road Type Speeds... 70

Table 6. Summary of All Road Type Traffic Volumes... 71

Table 7. Rural Interstate Crash Trend Results ... 75

Table 8. Comparison of Parallel and Interstate Routes ... 79

Table 9. Counter Naming Convention ... 103

(8)

List of Figures

Figure 1. Nationwide Fatal Crash Frequency by Roadway Function Class... 7

Figure 2. Nationwide Fatality Rate ... 8

Figure 3. Hypothesized Relationship between the Relative Change in Speed Limit Ranges and Safety... 13

Figure 4. Iowa Rural and Urban Interstate... 20

Figure 5. Iowa Rural Expressways ... 22

Figure 6. Iowa Rural Other Primary Highways ... 23

Figure 7. Parallel Routes... 24

Figure 8. Market Penetration of Rural Interstate Electronic Speeding Citations ... 32

Figure 9. Rural Interstate Fatal Crash Frequency ... 34

Figure 10. Rural Interstate Fatal Crash Frequency for Similar Periods ... 35

Figure 11. Rural Interstate Fatal and Major Injury Crash Frequency ... 35

Figure 12. Rural Interstate Fatal and Major Injury Crash Frequency for Similar Periods ... 36

Figure 13. Rural Interstate All Crash Frequency ... 36

Figure 14. Rural Interstate All Crash Frequency for Similar Periods ... 37

Figure 15. Rural Interstate Average and 85th Percentile Speeds ... 37

Figure 16. Percent of Vehicles Exceeding the Speed Limit by 10mph on the Rural Interstate ... 38

Figure 17. Rural Interstate Daytime and Nighttime Average Speeds ... 38

Figure 18. Rural Interstate Daytime and Nighttime 85th Percentile Speeds... 39

Figure 19. Percent of Daytime and Nighttime Vehicles Exceeding the Speed Limit by 10mph on the Rural Interstate ... 39

Figure 20. Rural Interstate Daytime Fatal Crash Frequency... 40

Figure 21. Rural Interstate Daytime Fatal Crash Frequency for Similar Periods... 41

Figure 22. Rural Interstate Daytime Fatal and Major Injury Crash Frequency... 41

Figure 23. Rural Interstate Daytime Fatal and Major Injury Crash Frequency for Similar Periods ... 42

Figure 24. Rural Interstate Daytime All Crash Frequency... 42

Figure 25. Rural Interstate Daytime All Crash Frequency for Similar Periods ... 43

Figure 26. Rural Interstate Daytime Fatal and Major Injury Crash Frequency and Average Speeds ... 43

Figure 27. Rural Interstate Nighttime Fatal Crash Frequency ... 44

Figure 28. Rural Interstate Nighttime Fatal Crash Frequency for Similar Periods ... 44

Figure 29. Rural Interstate Nighttime Fatal and Major Injury Crash Frequency ... 45

Figure 30. Rural Interstate Nighttime Fatal and Major Injury Crash Frequency for Similar Periods ... 45

Figure 31. Rural Interstate Nighttime All Crash Frequency ... 46

Figure 32. Rural Interstate Nighttime All Crash Frequency for Similar Periods ... 46

Figure 33. Rural Interstate Nighttime Fatal and Major Injury Crash Frequency and Average Speeds... 47

Figure 34. Urban Interstates Fatal Crash Frequency... 48

Figure 35. Urban Interstates Fatal Crash Frequency for Similar Periods... 48

Figure 36. Urban Interstates Fatal and Major Injury Crash Frequency... 49

Figure 37. Urban Interstates Fatal and Major Injury Crash Frequency for Similar Periods ... 49

Figure 38. Urban Interstates All Crash Frequency... 50

Figure 39. Urban Interstates All Crash Frequency for Similar Periods ... 50

Figure 40. 55-mph Urban Interstate Segments Average and 85th Percentile Speeds ... 51

Figure 41. 60-mph Urban Interstate Segments Average and 85th Percentile Speeds ... 51

Figure 42. Rural Expressway Fatal Crash Frequency ... 52

Figure 43. Rural Expressway Fatal Crash Frequency for Similar Periods... 52

Figure 44. Rural Expressway Fatal and Major Injury Crash Frequency... 53

Figure 45. Rural Expressway Fatal and Major Injury Crash Frequency for Similar Periods... 53

Figure 46. Rural Expressway All Crash Frequency... 54

Figure 47. Rural Expressway All Crash Frequency for Similar Periods... 54

Figure 48. Rural Expressways Average and 85th Percentile Speeds ... 55

Figure 49. Rural Other Primary Highway Fatal Crash Frequency... 56

Figure 50. Rural Other Primary Highway Fatal Crash Frequency for Similar Periods... 56

(9)

Figure 52. Rural Other Primary Highway Fatal and Major Injury Crash Frequency for Similar Periods... 57

Figure 53. Rural Other Primary Highway All Crash Frequency... 58

Figure 54. Rural Other Primary Highway All Crash Frequency for Similar Periods ... 58

Figure 55. Rural Other Primary Highway Average and 85th Percentile Speeds ... 59

Figure 56. Rural Non-Primary Roads Fatal Crash Frequency ... 59

Figure 57. Rural Non-Primary Roads Fatal Crash Frequency For Similar Periods ... 60

Figure 58. Rural Non-Primary Fatal and Major Injury Crash Frequency ... 60

Figure 59. Rural Non-Primary Fatal and Major Injury Crash Frequency ... 61

Figure 60. Rural Non-Primary All Crash Frequency ... 61

Figure 61. Rural Non-Primary All Crash Frequency for Similar Periods... 62

Figure 62. Rural Non-Primary Average and 85th Percentile Speeds... 62

Figure 63. All Rural Roads Fatal Crash Frequency ... 63

Figure 64. All Rural Roads Fatal Crash Frequency for Similar Periods ... 63

Figure 65. All Rural Roads Fatal and Major Injury Crash Frequency ... 64

Figure 66. All Rural Roads Fatal and Major Injury Crash Frequency for Similar Periods... 64

Figure 67. All Rural Roads All Crash Frequency ... 65

Figure 68. All Rural Roads All Crash Frequency for Similar Periods... 65

Figure 69. Percent Change in VMT by System ... 71

Figure 70. Relationship of Crashes to Traffic Volume (Qin et al., 2006)... 73

Figure 71. Rural Interstate Fatal Crash Frequency Trend Analysis ... 73

Figure 72. Rural Interstate Fatal and Major Injury Crash Frequency Trend Analysis ... 74

Figure 73. Rural Interstate All Crash Frequency Trend Analysis ... 74

Figure 74. Parallel Routes ATR Sites ... 76

Figure 75. Comparison of US-65 and I-35 Speeds ... 77

Figure 76. Comparison of IA-92 and I-80 Speeds ... 77

Figure 77. Ratio of Rural Interstate Electronic Speeding Citations to All Rural Interstate Electronic Citations Reported by the Iowa State Patrol ... 80

Figure 78. Rural Interstate ADT and Retail Gasoline Price by Time ... 81

Figure 79. Ratio of Rural Interstate ADT and Retail Gasoline Price by Time... 82

Figure 80. Rural Interstate Average Speeds and Retail Gasoline Price by Time ... 82

Figure 81. Location of Speed Adaptation Study Sites ... 98

Figure 82. Study Site 190th Street ... 99

Figure 83. Study Site 380th Street ... 99

Figure 84. Study Site IA-210 ... 100

Figure 85. Layout of the ATRs ... 102

Figure 86. 190th Street Lane One Vehicle Speeds Distribution at Counter 0... 104

Figure 87. 380th Street Lane One Vehicle Speeds Distribution at Counter 0... 104

Figure 88. IA-210 Lane One Vehicle Speeds Distribution at Counter 0... 105

Figure 89. Criteria for Experimental Vehicle Range of Arrival Times Criteria... 109

Figure 90. 190th Street All Vehicle Speeds Comparison... 112

Figure 91. 190th Street Passenger Vehicle Speeds Comparison... 113

Figure 92. 190th Street All PM Peak Period Vehicle Speeds Comparison... 113

Figure 93. 380th Street All Vehicle Speeds Comparison... 114

Figure 94. 380th Street Passenger Vehicle Speeds Comparison... 115

Figure 95. 380th Street All PM Peak Period Vehicle Speeds Comparison... 116

Figure 96. IA-210 All Vehicle Speeds Comparison ... 117

Figure 97. IA-210 Passenger Vehicle Speeds Comparison... 118

(10)

Acknowledgments

First and foremost I would like to thank my wife for the support she has given me as I worked on my graduate degree at Iowa State University. Her encouragement is a significant reason for my success. I would also like to thank my advisor, Dr. Reginald Souleyrette for his patience and willingness to help me through to a successful completion.

I would like to thank the 70-mph Steering committee and my graduate Program of Study committee for their advice and ideas regarding how to approach this project. I would like to thank Zach Hans for helping me learn and expand my knowledge on several technical aspects as I worked through my graduate program. He contributed a significant amount of expertise to forming the technical foundations for this project. I would like to thank Dan Gieseman for his willingness to answer technical questions and supplying some of the data used in this project. I would like to thank Brian Carlson from the Iowa Department of

Transportation for demonstrating generosity with his time and willingness to provide data for this project. I would like to thank LaDon Jones who created the computer programs needed for the data analysis. I would like to thank Lt. Larry Grant and Jim Saunders from the Iowa Department of Public Safety for their willingness to provide important information regarding the traffic citation data.

Additional thanks go to many of my peers. Specifically, I would like to thank Justin Jackson as he patiently helped me to learn GIS. I would like to thank Eric Fitzsimmons for contributing his personal time and expertise with much of the data collection. Additional thanks regarding data collection also go to Greg Karssen and Laurel Newman.

(11)

Abstract

On July 1, 2005, the speed limit on the rural interstates in Iowa was increased from 65-mph to 70-mph. This research first conducted a before and after study on the rural interstate and other facilities to study the effects on safety performance in Iowa due to this speed limit change. It explored the impact of the speed limit change on two effects known as the “speed adaptation” and “spillover effect.” Research was also conducted on traffic

citations issued on the rural interstate because citations may be a surrogate measure for highway safety. Finally, research was conducted on the recent increase in the retail price of gasoline and its effect on driver behavior. The rural interstates reported an increase in fatal crashes by 37.9 percent. No spillover effect in terms of crashes, speeds and volume were observed on other road types. Finally, no speed adaptation effect was observed in rural Iowa.

(12)

Chapter 1. General Introduction

1.0 Introduction

One of the essential components in providing safe roads is the speed limit. However, speed limit policy continues to be controversial. On July 1, 2005, the speed limit on the rural interstates in Iowa was increased from 65-mph to 70-mph. This change had been long considered by policy makers and the Iowa Department of Transportation (DOT) and was the subject of lively debate. Of concern was the impact the speed limit change on the rural interstate had on its safety performance and whether this change negatively affected other facilities (spillover) in terms of crashes and speeds. Additionally, the rural interstate speed limit introduced a 15-mph speed limit differential to rural primary highways that intersected and contained access with the rural interstate. This speed limit differential may have induced or augmented an effect known as the speed adaptation effect.

This research first examined crash performance on and off-system. It also explored the impact of the speed limit change on an effect known as the “spillover effect” to determine if increasing the speed limit on the rural interstates negatively affected other systems in terms of crashes or speeds. Because traffic citations reflect driver behavior, they may be a surrogate for highway safety. Research was then conducted on traffic citations issued before and after the speed limit change. If the recent increase in the retail price of gasoline did reduce the amount of travel, it may have partially masked any negative effects of the speed limit change. Research was conducted on the retail gasoline price and its corresponding effect on driver behavior in terms of the amount of travel. Finally, this research studied the speed adaptation effect in rural Iowa to determine if this effect exits and over what distance.

(13)

2.0 Thesis Organization

This thesis is divided into four chapters. This chapter provides a general introduction into the research topics. The second chapter reports on a before and after study of the safety effects of the rural interstate speed limit change. In addition to studying the rural interstates, other road types are also included in the analysis. These other road types include the urban interstates, rural expressways, rural other primary highways, and rural non-primary

highways. Further study was conducted on rural interstate traffic citations and the effect of the retail gasoline price on driver behavior.

The third chapter reports on a study on the driver adaptation effect in rural Iowa. It examines whether the change in the rural interstate speed limit has produced an immediate effect in terms of higher speeds on rural county and state highways intersecting the rural interstates. This study attempts to determine if this effect exits in rural Iowa and if so, for what distance. And finally, the fourth chapter provides a general conclusion of the findings of this research. Additionally, recommendations for further research into this topic are also provided.

(14)

Chapter 2. Iowa’s 70-mph Speed Limit: A Before and After Study

1.0 Introduction

One of the essential components of the highway system is the speed limit. Agent et al. (1998) states: “Appropriate speed limits are necessary to ensure a reasonable level of safe and efficient travel on highways and streets.” However, speed limit policy continues to be controversial. The debate centers on finding a proper balance between safety and efficiency of the highway system.

The United States’ economy is largely dependent on the transportation infrastructure and its ability to efficiently and reliably move people and goods. Efficiency of an

uncongested transportation facility is primarily achieved by increasing the speed limit. This reduces travel time and reduces the time component of user costs. But this increase can adversely affect highway safety. Joksch (1993) suggests that as a highway’s speed limit increases, so does the risk of a crash resulting in a fatality. Because of the seriousness of the risks, it is very important to closely monitor the after effects of policy decisions to increase speed limits in the context of highway safety.

The state of Iowa increased the speed limit on its rural interstates from 65-mph to 70-mph on July 1, 2005. This change in the rural interstate speed limit provides an opportunity to study the effects of the speed limit change on highway safety, and driver behavior.

Because of the nature of the highway system, numerous factors are continuously interacting. One challenge in studying the safety effects due to speed limit changes is isolating these factors. Kockelman (2006) reported that these factors include demographic changes, changes in the level, pattern, distribution, scheduling and purpose of travel, infrastructure

(15)

laws, alcohol laws, driver education and public safety campaigns, police enforcement, weather, and other secular trends. This report describes a preliminary before and after study on the safety effects of the rural interstate speed limit change.

2.0 Review of Literature

2.1 Determination of Speed Limits

When determining speed limit policy, decisions about posted speed limits should consider what most drivers would deem reasonable. Reasonable, as used in this context, refers to the speed at which drivers feel confident and safe in handling their vehicle. Najjar et al. (2000) states “Unrealistic posted speed limits generally reduce driver’s compliance rate. In addition, the number of accidents, related injuries and fatality rates may increase in these situations.” It would not be reasonable from a driver’s perspective to post a 45-mph speed limit on an interstate because drivers know that traveling at higher speeds is reasonable due to previous experience and personal judgment. Thus, the imposition of an unreasonable speed limit may then produce a higher rate of non-compliance which may result in reducing the overall safety prompting there to be greater variation in speeds selected by drivers. In the past, speed limits have not necessarily reflected the design speeds of certain roadway functions, mainly interstates. This apparent disparity between safety and efficiency eventually led to the debate and relaxation of speed limit policy in the United States.

2.2 History of Speed Limits

2.2.1 National History

In the United States, speed limit laws date to 1901 and traditionally have been left to the state’s authority to determine (Baum et al. 1989). The first federally regulated speed limit

(16)

was enacted during World War II to conserve fuel and rubber for the war effort. It was at this time that a national speed limit was set at 35-mph. After the war ended, this national speed limit was repealed. Most states then established posted speed limits of 65 and 70-mph on the United States highways (Garber and Graham, 1990). Then, in response to the 1973 Arab oil embargo, Congress enacted the National Maximum Speed Limit (NMSL) of 55-mph. In addition to saving fuel, a reduction in the number of highway fatalities was observed after the NMSL was enacted. Nationwide, 54,052 fatalities in 1973 were followed by 45,196 fatalities in 1974. Soon after, Congress made the NMSL permanent. This law also required that states certify that they were enforcing the NMSL (National Highway Traffic Safety Administration, 1998). To enforce the NMSL, the United States Department of Transportation (DOT) would withhold federal highway funds from states not meeting the new speed limit requirements (Garber and Graham, 1990). Following in 1978, the Surface Transportation Assistance Act required that states report the percentage of drivers exceeding the NMSL (National Highway Traffic Safety Administration, 1998).

The Surface Transportation and Uniform Relocation Assistance Act (STURA Act) was passed by Congress in 1987. This act relaxed the 55-mph speed limit on the nation’s interstates and allowed the states to raise the speed limit to 65-mph on rural interstates. In 1987, thirty-eight states increased their rural interstate speed limit followed by two additional states in 1988 resulting in about 90 percent of the nation’s interstate highways posting a speed limit of 65-mph. Most recently, the National Highway System Designation Act (NHS Act) of 1995 repealed the NMSL, returning complete authority of establishing speed limits to the states. By the end of 1996, thirty-two states had again raised their speed limits on various roadways (National Highway Traffic Safety Administration, 1998).

(17)

2.2.2 Iowa History

The earliest recorded law relating to the speed limit in Iowa dates to 1929 in which a statewide speed limit of reasonable and proper was established. The war effort of World War II created the need to set the maximum speed limit at a 35-mph which helped to conserve rubber and gasoline. At the end of the war, the speed limit was returned to the previous law of reasonable and proper in 1945. In 1957, a nighttime speed limit of 60-mph was

established. Soon after in 1959, the reasonable and proper speed limit was abolished. As a replacement of the reasonable and proper limit, each highway system was assigned a speed limit. Specifically, the interstate highways were assigned a 75-mph daytime and 65-mph nighttime speed limit, and primary highways were assigned a 70-mph daytime and 60-mph nighttime speed limit. No information was available on secondary roads (Crouch, 2006).

Because of the oil-embargo, the statewide speed limit was temporarily lowered to 55-mph in 1974 which was made permanent soon after in 1975. Following the passage of the STURA Act, Iowa increased its rural interstate speed limit to 65-mph in 1987, but retained the 55-mph speed limit for two-lane primary roads. Soon after the repeal of the NMSL in 1995, the speed limit of four lane divided highways were increased to 65-mph. Most recently, Iowa increased the speed limit on the rural interstate to 70-mph on July 1, 2005 (Crouch, 2006).

2.3 National Statistics Overview

Each year, approximately 40,000 to 45,000 highway fatalities occur in the United States. In 2005, the national fatality rate was reported as 1.45 fatalities per 100 million vehicle miles traveled (National Highway Traffic Safety Administration, 2005). By functional class, interstates experience the lowest number of fatal crashes among any

(18)

roadway functional class as shown in Figure 1. However, crashes on the interstate have the potential to be more severe because of the high speeds (Joksch, 1993). Arterials exhibit the largest number of fatal crashes.

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Year F a ta l C ra s h F re q u e n c y Interstate Arterial Collector Local

Figure 1. Nationwide Fatal Crash Frequency by Roadway Function Class Source: FARS (Accessed 9/19/06)

The National Highway Traffic Safety Administration (2004, 2005) Annual Report on Traffic Safety Facts provides data on the nationwide fatality rate beginning in 1966 as shown in Figure 2. The three major congressional acts related to speed limit policy are placed on Figure 2 to illustrate their position in time relative to the national fatality rate. The referenced reports did not include the fatality rate from 1967 to 1969, therefore no numbers are reported in Figure 2 for these years.

(19)

0.00 1.00 2.00 3.00 4.00 5.00 6.00 1 9 6 6 1 9 6 7 1 9 6 8 1 9 6 9 1 9 7 0 1 9 7 5 1 9 7 6 1 9 7 7 1 9 7 8 1 9 7 9 1 9 8 0 1 9 8 1 1 9 8 2 1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1 9 8 7 1 9 8 8 1 9 8 9 1 9 9 0 1 9 9 1 1 9 9 2 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 Year F a ta li ty R a te (p e r 1 0 0 m il li o n v e h ic le m il e s t ra v e le d )

Figure 2. Nationwide Fatality Rate Source: NHTSA 2004 and 2005 Annual Report

The fatality rate decreased dramatically from 1970 to 1975. Then the fatality rate increased slightly until 1979 after which the rate gradually decreased until 2004. In 2005, the fatality rate increased. A dramatic decrease in the fatality rate did occur after the NMSL was set at 55-mph. Following the 1987 STURA Act and 1995 NHS Act, the fatality rate

continued to decrease. There was concern that fatalities would increase after the relaxation of the speed limit in 1987 and the complete repeal of the NMSL in 1995; but, this has not been observed. However, one may counterfactually hypothesize that the nationwide fatality rate would have been lower had the NMSL not been abolished. Although the acts of 1987 and 1995 coincide with continued decreases in the fatality rates, there are many other factors that contributed to improving highway safety, such as more stringent seat-belt laws, safety improvements in vehicles, public education programs, and better emergency response.

1973 NMSL Enacted: National Speed Limit

of 55-mph

1995 NHS Act: Repealed NMSL 1987 STURA Act:

Allowed Maximum Rural Interstate Speed Limit of

(20)

2.4 Research Methodology Controversy

Speed limit policy is a very controversial topic. Part of this controversy centers on the research methodology of speed limit safety studies. In an article supporting a possible safety benefit of the rural interstate 65-mph speed limit of the late 1980s, Lave and Elias (1994) hypothesized that the NMSL of 55-mph resulted in a misallocation of police resources. When the NMSL was made permanent in 1974, states were financially pressured to place their speed limits at 55-mph and to report the proportion of drivers exceeding the speed limit. This required additional enforcement by police departments thus reducing the amount of time spent patrolling other facilities such as the more dangerous county and state highways. Because of a resulting higher concentration of police patrols on the interstate, it was argued that this enforcement would also lower the interstate crash rate producing a rate much lower than it would have been if the enforcement had remained the same. Lave (1995), quotes a member of the International Association of Chiefs of Police as stating “[Federal financial sanctions] force the over-concentration of limited resources for the express purpose of attaining compliance rather than application of resources in a manner most effectively

enhancing total highway safety…” Thus, Lave and Elias suggested that after the relaxation of the rural interstate speed limit allowing a 65-mph speed limit and the requirement to patrol the rural interstates, police departments could shift their resources to patrolling other facilities thereby enhancing highway safety.

Lave and Elias (1994) also hypothesized an overall safety benefit of the 65-mph rural interstate speed limit. The hypothesis stated that because rural interstates, state highways, and county roads were all posted at a 55-mph speed limit before the enactment of the STURA Act of 1987 allowing the speed limit to be increased to 65-mph on the rural interstates,

(21)

drivers that chose to speed would chose to not drive on the interstate but instead drive on county and state highways which would lower their chance of being caught speeding due to the increased enforcement on the rural interstates. Additionally, drivers may have chose to drive on the county and state highways since they could provide a possible benefit of shortened travel times between destinations due to a more direct route to their destinations. When the rural intestate 65-mph speed limit was reinstated, the change could have produced a shift in traffic volumes from two-lane county and state highways to safer rural interstate facilities because drivers may have determined it would shorten travel time. In support of their hypothesis, Lave and Elias (1994) reported that states which raised their speed limits to 65-mph on rural interstates in 1987 had an overall drop in the statewide fatality rate of 6.15 percent for the years of 1987 and 1988. For states which did not raise their speed limit, they were reported to have an overall drop in the statewide fatality rate of only 2.62 percent for the same years.

Articles opposing the hypothesis’ purported by Lave and Elias were written following the increase in speed limit during 1987 and 1995. Baum et al. (1989) and Baum et al. (1991) reported that states which increased their speed limit to 65-mph on rural interstates found there was an increase in fatalities on the rural interstates while there was no similar trend for states that did not raise speed limits on their rural interstates. In a comparison of the crash history for states which raised their rural interstate speed limit to at least 70-mph in

1995/1996 to those that did not, Farmer et al. (1997) reported that those states which raised the speed limit experienced a 16 percent increase in the number of fatalities on their interstate system following the change versus a 4 percent increase for those states that did not increase their interstate speed limits. In a later article, Farmer et al. (1999) quoted the Insurance

(22)

Institute for Highway Safety president as stating “It’s clear from this study that the current round of speed limit increases, like increases on rural interstates in the 1980s, is costing hundreds of lives per year”

As a rebuttal to the studies reporting that an increase in the speed limit increases the number of fatalities, it was argued by Lave and Elias (1994) that those studies only looked at the number of fatalities on those highways which were affected by the change. It was

suggested that to study the safety effects of a speed limit change, it is necessary to assess the impact on the entire or statewide system by using fatality rates.

2.5 Summary of Speed Limit Safety Studies

An extensive review was conducted on several statewide studies that analyzed the safety effects of increasing interstate speed limits. These studies included analysis periods which covered the speed limit changes of the 1987 STURA Act and repeal of the NMSL in 1995. Table 1 summarizes some of the findings that were reported.

Most of the studies did report an increase in various crash severities after an increase in the speed limit on their respective interstate facilities. However, it was reported in Kansas, Oklahoma, and Utah that there was no adverse effects on safety due to the increase in the speed limit on their interstates. Each of these states increased their rural interstate speed limits to 70 or 75-mph.

(23)

Table 1. Summary of State Interstate Speed Limit Studies Rural Interstate Speed Limit State Author(s) Before After Results Illinois Sidhu, 1990 55 65

Fatal, injury and property damage crashes all experienced an increase in the after period for rural interstates. Only property damage crashes were found to have a statistically significant increase using a Chi-Squared test.

Illinois Rock, 1995 55 65

A statistically significant increase in all crashes, fatalities, and injuries was found on interstates posted at 65mph for the after period by using a t-test. An ARIMA model calculated an additional 345 crashes, 15 fatalities, and 150 injuries resulted on rural highways due to the 65mph speed limit

Iowa Ledolter and

Chan, 1996 55 65

Found an 82% increase in fatal crashes on the rural interstates with a before period of 1983-1986 (56 fatal crashes) and an after period of 1988-1991 (102 fatal crashes). Increase was found to be statistically significant under a Poisson assumption.

Iowa Raju et al., 1998 55 65

Using a Bayesian approach, an additional four fatal crashes occurred per quarter on the rural interstate due to the speed limit change.

Kansas Najjar et al., 2000 65 70 There was no significant increase in the after period for all crash, fatal crash, and fatality rates.

Kentucky Agent et al., 1998 Various Various

55mph interstate segments had a fatal and injury crash rate of 0.39 crashes/161MVKM and 30 crashes/161MVKM,

respectively. 65-mph interstate segments had a fatal and injury crash rate of 0.44 crashes/161MVKM and 23

crashes/161MVKM, respectively. Michigan1 Binkowski et al.,

1998 65 70

No spillover effect, in terms of vehicle speeds, was found on facilities located near interstates in which there was an increase in the speed limit. Crash data was not analyzed.

Minnesota Minnesota DOT,

2007 65 70

A 70% increase (32% increase adjusted by vmt) in fatal crashes was observed on the rural interstates after the speed limit was increased. The study used a 5 year before and 5 year after period.

New Mexico Gallaher et al.,

1989 55 65

A statistically significant increase was found in fatal crashes, fatalities, and fatal single-vehicle crashes under the assumption of a Poisson distribution. No change was found for multi-vehicle fatal crashes.

North Carolina

Renski et al.,

1998 65 70

Interstate facilities that had an increase in the speed limit of 10mph were found to have more risk of increase crash severity than those that only had an increase in the speed limit of 5mph.

Oklahoma2 Oklahoma DOT,

1998 65 75

No statistically significant increase in overall crash frequencies or crash rates were observed on the rural interstates.

Additionally, no statistically significant change occurred in crash severity.

Texas Brackett and Ball,

1990 55 65

A statistically significant increase was found in injury, property damage, total and serious crashes.

Utah3 Vernon et al.,

2004 65 75

No increase in the total, fatal, injury crash rates were observed on rural interstates in the after period.

Washington State

Ossiander and

Cummings, 2002 55 65

Found the fatal crash rate on rural interstates was 110% higher than it would have been had the speed limit change not been changed. The overall crash rate on all interstates showed little change.

1. Trucks remained at 55-mph

2. Non-turnpike rural interstates were posted at 70-mph 3. Small proportion of interstates increased to 70-mph

(24)

Kockelman (2006) suggested that in cases of an increase in a speed limit from a lower range (55-mph to 65-mph), there may be a greater change in the number of crashes in the after period versus a speed limit change from a higher range (65-mph to 75-mph). It was suggested that for speed limit changes encompassing higher ranges, drivers may already be cautious due to the already existing high speed environment. Figure 3 illustrates this hypothesized relationship between the differences of the range of the speed limit change to the probability of a fatality.

Figure 3. Hypothesized Relationship between the Relative Change in Speed Limit Ranges and Safety Source: Kockelman, 2006

2.6 Spillover Effect

The transportation system can be considered an open system. A change in the

operating conditions at one location of the transportation system may have an effect on other portions of the system as well. Changing a parameter at one location in the transportation system and the effect it has on other portions of that system has been named the “spillover effect” (Binkowski et al., 1998; Kockelman, 2006; Ledolter and Chan, 1996; Pant et al.,

Low Range (55-65mph) High Range (65-75mph C h a n g e in P ro b a b il it y o f F a ta li ti es

(25)

1992; Rock, 1995; Srinivasan, 2002). Kockelman (2006) provides a definition of the spillover effect with respect to vehicle speeds as “…the impact that a speed limit change on one road may have on parallel facilities.” It was suggested that urban areas may also be more susceptible to spillover effects than rural areas because of networks that are much denser. Three areas in which the spillover effect may be active are vehicle speeds, traffic volumes and crashes.

Garber and Graham (1990), in a state-by-state study of the effects of the 65-mph speed limit on interstate highways, summarized two opposing hypotheses known as “traffic diversion” and “speed spillover”. In the context of increasing the speed limit on rural interstates, traffic diversion hypothesizes that traffic would shift to the rural interstates thereby decreasing the fatalities on rural non-interstate highways. In the same context, speed spillover hypothesizes that fatalities would increase on rural non-interstate highways because of an increase in the speed limit on the rural interstates. Srinivasan (2002) summarized the speed spillover effect in such a way that if it were to exist “…it can lead to increase in average speeds on roads where the speed limit was not raised and are not designed to handle high-speed traffic.” This effect is characterized by drivers that “…get in the habit of driving faster and do so even on roads that have not had their speed limits raised” and those who might “…fail to slow down upon exiting a rural interstate and continuing their journey on roads with lower speed limits” (Garber and Graham, 1990).

2.6.1 Speed Spillover Effect

A study conducted by Ledolter and Chan (1996) evaluated the impact of increasing the speed limit on rural interstates in Iowa from 55-mph to 65-mph. Rural interstates, urban interstates, rural primary, and rural secondary roads were examined in the study with only the

(26)

rural interstates experiencing a change in the speed limit. The rural interstates recorded an increase in vehicle speeds from 59-mph in the before period to 66-mph in the after period. For the other road classes that did not have a change in the speed limit, the average vehicle speeds increased by 1-mph. The speed spillover effect in this case was concluded to be “small.”

Binkowski et al. (1998) studied the effect on speeds and traffic volume due to the increase in the speed limit from 65-mph to 70-mph on rural freeways in Michigan. Part of the study tested for the presence of a speed spillover effect from road segments with an increased speed limit to those sections in which the speed limit remained the same. Roadway segments used as a control group included intercity, urban and recreational freeway, and rural two-lane highways. The posted speed limits for the control segments were 55 and 65-mph.

Experimental segments included intercity and recreational routes. The posted speed limit for all experimental segments was 65-mph during the before period and 70-mph for the after

period. The 50th and 85th percentile speeds were measured on the control and experimental

segments during the before and after periods. The control segments experienced an increase

in the 50th and 85th percentile speeds in the after period of 0.1 to 0.8-mph, while the largest

decrease observed was 0.3-mph. It was concluded that the control segments experienced no speed spillover effect.

Brown et al. (1990) conducted a case study of Alabama’s 65-mph rural interstate speed limit. A portion of the study tested if there was a spillover effect from roads that were posted at 65-mph to those posted at 55-mph. The first test studied segments of the interstate that retained the 55-mph speed limit. Sites were selected so that speed adaptation would not be a confounding factor. The second test studied segments of non-interstate roads with a

(27)

speed limit of 55-mph that were proximal to interstates with a 65-mph speed limit and that had access to the 65-mph interstate. The study found that the speed spillover effect onto 55-mph interstate and 55-55-mph non-interstate roads from 65-55-mph rural interstates was about the same with an increase of about 1-mph.

Godwin (1992) studied the effect of the 65-mph speed limit on safety for the entire United States. He found evidence that there was a speed spillover effect from 65-mph interstates to 55-mph rural interstates, but stated “The effect of speed spill-over to non-Interstate roads in both 55-mph and 65-mph states is uncertain.” McKnight and Klein (1990) compared states which increased the speed limit on their rural interstates to 65-mph to those states which retained the 55-mph speed limit. They determined that in states which increased the speed limit on rural interstates to 65-mph, the percentage of vehicles exceeding the speed limit on 65-mph and 55-mph roads were both found to be significant. Garber and Graham (1990) reviewed data from states which increased the speed limit on rural interstate highways to 65-mph. Preliminary results suggested that the 65-mph speed limit did have an effect on rural non-interstates in terms of traffic diversion and speed spillover, but speed spillover was determined to have a larger effect than traffic diversion.

In a continuation of their 1985 study of speed adaptation, Casey and Lund (1992) suggested that by increasing the speed limit on some roads, it affected vehicle speeds on other roads up to two hours of driving time away. This may indicate that a change in the speed limit at one location in the transportation network can have far reaching effects throughout the network.

(28)

2.6.2 Traffic Diversion Effect

Rock (1995) studied the impact of increasing the speed limit on rural interstate and limited access-highways from 55-mph to 65-mph in Illinois. Vehicle miles traveled (VMT) for rural interstates increased significantly above the trend of the four years prior to the speed limit change. For non-interstate rural highways, VMT decreased relative to both the trend of the previous four years prior the speed limit change and in absolute terms. It was concluded that “…some traffic was diverted from highways with a 55-mph speed limit to 65-mph highways.”, and the 65-mph highways may have actually “generated new traffic.” It was also suggested that a consequence of this traffic diversion was that it produced a speed spillover onto 55-mph highways.

In Alabama, Brown et al. (1990) calculated a ratio of the average daily traffic (ADT) observed on rural interstates to non-interstate principal arterials. Before the speed limit change, the ratios ranged from 1.03 to 1.26. Following the speed limit change, the ratio was calculated as 1.49. It was concluded that “…shifts to the interstates are occurring from the non-interstate categories…” And in a study of the highway safety effects due to the 65-mph speed limit in Indiana, McCarthy (1991) also suggested that there was a shift in traffic from “lower-speed roads” to the interstates.

2.6.3 Crash Spillover Effect

In Iowa, Ledolter and Chan (1996) found that rural interstates experienced an

increase in the number of fatal crashes by 82.1 percent in the after period. Roads that did not have a change in the speed limit, urban interstates, rural primary and rural secondary roads, experienced a change in the number of fatal crashes by -18.2, 8.1 and 1.5 percent,

(29)

speed limit had a large effect on rural interstates, but the effect on urban interstates was found to be “essentially zero.” Major injury crashes on rural interstates increased while they decreased on urban interstates, rural primary and rural secondary roads. Further analysis estimated that the increase in the speed limit produced an additional two fatal crashes on rural interstates, six fatal crashes on rural primary roads and four fatal crashes on rural secondary roads per quarter. A more conservative estimate calculated additional fatal crashes occurred on rural interstates, primary and secondary roads per quarter year.

A crash model developed by Rock (1995) suggested that on rural highways in Illinois, an additional 345 crashes, 15 fatalities, and 150 injuries occurred because of the speed limit change. The study cited higher speeds, increase in speed variance, traffic diversion, traffic generation and speed spillover as possible factors in producing additional crashes and injuries.

In Alabama, Brown et al. (1990) found that interstates with a 55-mph speed limit, property damage only crashes (PDO) increased significantly in the after period, but there were no significant increase in fatal and injury crashes. Non-interstate roads with a speed limit of 55-mph and proximal to the 65-mph interstates did not experience any significant increase for any of the different crash severities.

Pant et al. (1992) studied the effects of increasing the rural interstate speed limit in Ohio from 55-mph to 65-mph. Three types of roads were studied. They were rural interstates posted at 65-mph, rural interstates posted at mph, and rural non-interstates posted at 55-mph. For rural interstates with a 55-mph speed limit, mean fatality rates increased

significantly in the after period. However, when adjusted for “normal” and “adverse” weather conditions, no significant difference was found. Injury and PDO crashes decreased in the

(30)

after period. For non-interstate highways with a 55-mph speed limit, there was no significant difference in the mean fatal crash rates. Similar to the rural interstate segments posted at 55-mph, injury and PDO crashes decreased in the after period. It was concluded that there were no negative consequences from the spillover effect.

In a state by state comparison, McKnight and Klein (1990) reported that roads posted at 55-mph and 65-mph had a significant increase in fatal crashes from the before to after period. Garber and Graham (1990) provided preliminary results that indicated that the 65-mph speed limit was affecting the number of fatalities on rural non-interstates in addition to the rural interstates.

3.0 Methodology

This report analyzed the effects of the rural interstate speed limit change through a

before and after study. Crash data were obtained for the period January 1, 2003 to December 31, 2006. This provided up to 30 months for the before period and up to 18 months for the after period, for crashes, vehicle speeds and traffic volume. The analysis period for crashes and traffic volume were the same, but was different for speeds.

Because the increase in the speed limit occurred only on the rural interstates, other road types were analyzed to test for any type of spillover effect. Six road types were analyzed. These included:

• Rural Interstates

• Urban Interstates

• Rural Expressways

• Rural Other Primary Highways

• Rural Non-Primary Roads

(31)

3.1 Rural and Urban Interstates

The primary objective of this research was to study the safety effects of increasing the

speed limit on the rural interstates in Iowa. Figure 4 displays rural and urban interstates in

Iowa. The rural interstates consist of 622 miles of roadway while urban interstates consist of

143 miles. Urban interstate segments were defined as any segment located within one of

Iowa’s metropolitan planning organizations (MPO) planning areas. Any segment located

outside of an MPO planning area was deemed rural. The urban interstate segments were

located within the urban areas of Sioux City, Council Bluffs, Des Moines, Waterloo, Cedar Rapids, Iowa City and Davenport.

Figure 4. Iowa Rural and Urban Interstate

There are other ways urban interstate segments could be defined. For example, urban

(32)

posted on the rural interstate. Urban interstates could also be defined as any segment that is located within the corporate limits of a city with a specified population. The issue with such definitions in Iowa is that portions of the interstate segments posted with a lower speed limit

or located within corporate limits sometimes resemble rural interstate segments in terms of

the number of access points and urban development. As demonstrated by these examples,

any definition of urban interstate segments will ultimately consist of some subjective

elements.

3.2 Rural Expressways

For this research, rural expressways consisted of four lanes, any type of median (hard

surface without barrier, grass surface without barrier, hard surface with barrier, grass surface with barrier), under the jurisdiction of the Iowa DOT, and with at-grade intersections.

Because of their similarity in design standards to the interstate system, any spillover effect to non-interstate facilities may first be expected thereon. Crash, volume and speed data were collected and analyzed for all Iowa expressways.

In Iowa, two-lane primary highways may become four-lane expressways for a short distance, thereafter reverting to a two-lane primary highway. For this study, any road

segment which matched the definition of a rural expressway was defined as a rural

expressway. This definition included many short and long segments. The rural expressways

consist of 1,155 miles of roadway. Figure 5 displays the rural expressway road network as

(33)

Figure 5. Iowa Rural Expressways

3.3 Rural Other Primary Highways

As with rural expressways, a spillover effect may also be observed on rural other

primary highways. Rural other primary highways were defined as any road segment that is identified as part of the national or state highway system within Iowa, but is not considered

an interstate or expressway highway. The rural other primary highways consist of 5,885

miles of roadway. Figure 6 displays the road network for the rural other primary highway

(34)

Figure 6. Iowa Rural Other Primary Highways

3.4 Rural Non-Primary Roads

The road classification within Iowa that consists of the largest amount of roadway

miles is rural non-primary roads. Rural non-primary roads included all rural roads not

defined as national or state highways, rural expressways or rural interstates. These roads are

maintained by the various county agencies. The majority of these roads are aggregate

surfaced (gravel) roads. Rural non-primary roads consist of 90,040 miles of roadway.

3.5 Primary Parallel Routes

This research also investigated the spillover effect on primary parallel routes to the

interstate. Primary parallel routes consisted of any route that is part of the national or state

(35)

competing route to a rural interstate. The two primary parallel routes chosen for this study

are US-65, and IA-92. Figure 7 displays the primary parallel routes used for this study.

Figure 7. Parallel Routes

3.6 Rural Interstate Traffic Citations

With the passage of the 70-mph speed limit in Iowa, additional Iowa State Patrol enforcement was promised by the governor and legislature (Iowa Department of Public Safety, 2006). A change in the number of speeding citations could be considered as a surrogate for a change in the number of speed related crashes. Because the speed limit changed on the rural interstate, traffic citation data were only collected and studied for the rural interstate.

(36)

3.7 Effect of Retail Gasoline Price

The recent increase in the retail price of gasoline may have affected the travel

behaviors of drivers in Iowa. Because of the higher prices, drivers may tend to drive less and attempt to conserve fuel by driving at lower speeds than they otherwise would have.

Therefore, it is possible that the higher cost of gasoline may have indirectly “canceled” out

some effects of the increase of the rural interstate speed limit. To determine if the cost of

gasoline did have an effect on driver behavior and therefore offset some of the possible negative impacts of the speed limit change, the price of gasoline for the recent history in Iowa was collected. It was presumed that if drivers altered their behavior due to the higher cost of gasoline, this behavior would be most pronounced on a facility that consists of longer

trips, namely rural interstates.

4.0 Data Analysis

4.1 Crash Data

The Iowa DOT maintains an extensive crash database. Crash data are submitted to the Iowa DOT from various police agencies such as the Iowa State Patrol, county sheriff offices, and city police departments and are usually submitted by these various police agencies at different times. For example, some police agencies may provide crash data on a weekly basis while others may only provide data once a month. When the crash data are submitted to the Iowa DOT, the data undergoes an editing process. The crash data used for this study were downloaded from the Iowa DOT database on April 2, 2007.

The crash severities for Iowa crash data are aggregated by fatal, major injury, minor injury, possible/unknown, and PDO. The severity of a crash is defined by the worst injury of

(37)

the crash. The injury severities of those involved in a crash are estimated by an officer on the scene. In the officer’s reporting guide, definitions are provided to aid the officer in

identifying the severity of those involved in a crash. A crash is defined as fatal if any person involved in a crash died as a result of their injuries sustained from that crash within 30 days. A crash is defined as a major injury crash if any person’s injuries sustained from that crash prevents that person from walking, driving or continuing with normal activities that they were capable of before the crash. Other indications of a major injury included severe lacerations, broken or distorted limbs, skull, chest, or abdominal injuries, unconsciousness, and unable to leave the crash scene without assistance. A crash is defined as a minor injury crash if any person’s injuries are evident to those at the crash, but are not included in the definition of a fatal or major injury. Indications of a minor injury are lumps on a head, bruises, abrasions, and minor lacerations. A crash is defined as possible/unknown if a person involved in a crash reports or claims a personal injury sustained during that crash that is not included in the definition of fatal, major, and minor injuries. Indications of a

possible/unknown crash are momentary unconsciousness, claim of injuries that are not evident, limping, complaint of any pain, nausea, and hysteria. A crash is also defined as a possible/unknown crash if a reporting officer does not know if any injuries were sustained from the crash. The final crash severity classification is property damage only (PDO). Crashes are defined as PDO crashes if the crash only resulted in the damage to the vehicle(s) involved in that crash. The reporting threshold of PDO crashes is $1,000 (Iowa Department of Transportation, 2001).

The crash data used for this study include a variety of information about each crash. Each crash is assigned a geographical coordinate which can be mapped using a geographical

(38)

information system (GIS). Crashes determined to be within a specified distance of any facility of interest were assigned to that road type within ArcView GIS 3.3. The 2004 statewide road network most accurately reflects the geographical location of crashes that occurred during the analysis period of this research. A total of 3,261 out of 232,061 crashes (for the years of 2003 to 2006), or 1.4 percent, were not assigned a coordinate, thus they could not be located on a map and could not be used. Once the crash data were assigned to the appropriate road types, they were summarized by the number of crashes within the various severity levels on a monthly basis. The severity levels included in this analysis were fatal, fatal and major injury, and all crashes.

4.2 Speed and Volume Data

In conjunction with the crash data, speed and volume data were collected for each facility type from the Iowa DOT. The Iowa DOT maintains many permanent automatic traffic recorders (ATRs) throughout the state. Because of the existing structure of the ATR databases, it was necessary to create a computer program to extract the speed data. During discussions with the Iowa DOT, it was noted that the ATR database was changed during the summer of 2004. Because of this change, the analysis period for the speed data included

August, 2004 to December, 2006. The speed data were summarized by the average, 85th

percentile speeds, and the percent exceeding a given speed value threshold. Volume data were obtained from the monthly automatic traffic recorder reports provided on the Iowa DOT’s website and included the same analysis period as crash data.

4.3 Rural and Urban Interstate Crashes

Only crashes located on the interstate mainline would be considered for the analysis. Crashes located on the interstate ramps were assumed to be unrelated to any effect of the

(39)

speed limit change. Once the crashes were spatially assigned to the interstate, they were then

aggregated by whether they occurred on a rural or urban interstate segment. Rural interstate

crashes were additionally designated as daytime or nighttime.

The first step in spatially assigning crashes to the interstate was to select all interstate road segments from the 2004 statewide road network database. These selected road segments were placed into a mainline interstate road network. All other road segments not considered to be mainline interstate were selected and placed into a non-interstate mainline road

network. The two road networks were then joined to the master crash database and a distance attribute was calculated for each road type.

Two additional fields were added to the master crash database: absolute difference in distance and LOC (level of confidence). Absolute difference between the two distance fields was created for interstate and non-interstate road segments. The LOC field was used to identify the confidence in which a facility was assigned to a crash. A value of 1 stated that a crash was assigned to the facility of interest (in this case the interstate) with a high degree of confidence. A value of 5 stated that a crash was assigned to another facility (in this case all non-interstate mainline roads) with a high degree of confidence. A crash with a LOC value of 3 was assigned to the facility of interest, but with a lower confidence than a value of 1. Some crashes are clearly located on a facility of interest. Other crashes are clearly located on other facilities. However, some crashes are “in-between” the two types of facilities. This iterative process was therefore implemented to provide a systematic approach to assigning these remaining “in-between” crashes to their respective facility type. A summary of the LOC values is found in Table 2.

(40)

Table 2. Road Segment Assignment Scheme

LOC Assigned Facility 1 Facility of Interest 2 Facility of Interest 3 Facility of Interest 4 Other Facility 5 Other Facility

The crash assignment process consisted of a series of queries with various conditions. The distance values were reported in meters. The query conditions for assigning crashes to their respective road type and the order in which they were completed are:

First Query: OF FI D D ≤ and D_FI <25 LOC = 1 Second Query: OF FI D D > and DiffDist >50 LOC = 5 Third Query: OF FI D D ≤ LOC = 2 Fourth Query: OF FI D D > and DiffDist >5 LOC = 4 Fifth Query:

All remaining crashes LOC = 3

FI

D = Distance to facility of interest

OF

D = Distance to other facility

DiffDist= Absolute difference between distance to facility of interest and other facility

The second step of the crash assignment process was to run a query that selected the crashes occurring on the facility of interest with the condition of LOC equals 1, 2 or 3. After

(41)

the crash assignment process was completed, the crashes located on the interstate were defined as occurring either in a rural or urban area. An interstate crash located within 50

meters of any rural interstate segment as defined in section 3.1 was deemed rural. All other

interstate crashes were deemed urban by default.

4.4 Other Road Types

All expressway road segments located outside of corporate limits were defined as “rural expressways.” Non-rural expressway road segments were also selected using the 2004 statewide road network. Crashes were then assigned to their respective facility and analyzed by the same method as discussed in section 4.3.

All non-interstate, non-expressway primary road segments located outside corporate

limits were selected and defined as “rural other primary” road segments, again using the

2004 statewide road network. Crashes not previously assigned to the rural or urban

interstates, or rural expressways were then assigned to the closest rural other primary highway by the same method as discussed in section 4.3. All remaining crashes located

outside of any corporate limit were defined as “rural non-primary” crashes. Rural

expressways, rural other primary, and rural non-primary crashes were not analyzed for day/night safety performance.

Finally, to test if there was a shift in speeds or traffic volumes on primary parallel

routes, ATRs from primary parallel routes and corresponding interstates were selected for a

comparison analysis. For this purpose, IA-92 was paired with I-80 and US-65 was paired with I-35. Speed and traffic volume data were then compared. No crash data were analyzed in this step.

(42)

4.5 Daytime and Nighttime Rural Interstate Crashes

For this study, rural interstate crashes were defined as occurring during the day or

night based upon official sunrise and sunset times obtained from the United States Naval Observatory. Rather than entering each segment location, approximations were developed for sunrise and sunset times. Since the sunrise and sunset times can only be obtained for one location, it was necessary to decide which location would be used for Iowa. Because of the longitudinal width of the state of Iowa, the sunrise and sunset times are different for the east and west ends of the state. The difference between the sunrise and sunset times were

determined by selecting Davenport and Council Bluffs and obtaining their respective sunrise and sunset times for an arbitrary date of May 9, 2007. The sunrise and sunset times for Council Bluffs, Iowa were 6:12am and 8:29pm, respectively. The sunrise and sunset times for Davenport, Iowa were 5:50am and 8:08pm, respectively. The difference between the sunrise times is 22 minutes while the difference between sunset times is 21 minutes. Therefore, it was assumed that the sunrise and sunset times for a central location in Iowa would approximate the sunrise and sunset times for every location within the state. The central location selected was Ames, Iowa. The sunrise and sunset times were obtained for 2003 to 2006. To account for daylight savings time, the United States Naval Observatory noted that for days occurring during daylight savings time, one hour should be added to the times provided in the table.

Effective sunrise and sunset times defined for each day of the year were joined to the rural interstate crash database on the basis of the date. This join resulted in assigning the sunrise and sunset times for the day in which each crash occurred. To define crashes

(43)

must be less than the sunset time and greater than the sunrise time. All other crashes not selected were defined as occurring during the night by default.

4.6 Rural Interstate Traffic Citations

A manual defining the various types of traffic citations was provided by the Iowa State Patrol. A database consisting of all electronic traffic citations issued by the Iowa State Patrol was obtained through the Center for Transportation Research and Education (CTRE) at Iowa State University. The structure of the citation database allowed it to be queried and mapped in ArcView GIS 3.3. Citations were first assigned to the rural interstates with the method described in section 4.3. Speeding citations were then selected from the rural interstate traffic citation set.

Electronic traffic citations have only recently been put into use by the Iowa State Patrol. Over the last three years, the number of traffic citations has greatly increased as shown in Figure 8. 0 500 1000 1500 2000 2500 J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r 2003 2004 2005 2006 Time F re q u e n c y

(44)

As shown in Figure 8, there were few citations in the database in 2003. Soon after in 2004, the number of electronic citations began to increase dramatically. According to the Iowa State Patrol, the current (July, 2007) utility of the electronic citation database is approximately 80 percent but should be at 100 percent by December, 2007. Because the number of paper traffic citations issued by the Iowa State Patrol was not known, the relative electronic share could not be determined. Therefore, the number of electronic speeding citations could not be compared before and after the change in the speed limit. However, the ratio between electronic speeding citations to the total number of electronic citations on the rural interstate was calculated and plotted for 2004 to 2006. An increase in the ratio of

speeding citations to all other citations after the speed limit change could indicate an increase in the number of speeding related citations issued by the Iowa State Patrol.

4.7 Effect of the Retail Price of Gasoline

Data for the retail price of gasoline were obtained from United States Department of Energy (Energy Information Administration, 2007). The gasoline formulation selected for this analysis was regular grade gasoline, as sold through retail outlets. Prices were recorded on a monthly basis. Two charts were created that displayed the retail gasoline price plotted

with the rural interstate ADT and average speed over time. The analysis period was selected

as January, 2002 to December, 2006 coincident with the availability of ATR reports.

5.0 Results

For each of the road types that were analyzed, two crash charts are displayed for each crash severity. The first crash chart displays January, 2003 to December, 2006 crash data. The shaded boxes cover portions of the observed time period to allow similar months to be

(45)

compared in the before and after time periods, as only 18 months of crash data are available after the speed limit change. For example, see Figure 9. A second chart displays only before and after periods over similar time periods to facilitate comparison. For example, see Figure 10.

5.1 Rural Interstates

Figures 9 through 14 display the rural interstate crash frequency before and after the

speed limit change for fatal, fatal and major, and all crashes. Figure 9 displays the rural

interstate fatal crash frequency. The two months with the largest fatal crash frequency are February, 2005 and November, 2005. The number of months not recording fatal crashes before the speed limit change was four while only one month did not record fatal crashes in

the after period. Figure 10 displays the rural interstate fatal crash frequency for similar time

periods before and after the speed limit change.

0 1 2 3 4 5 6 J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r J a n u a ry F e b ru a ry M a rc h A p ri l M a y J u n e J u ly A u g u s t S e p te m b e r O c to b e r N o v e m b e r D e c e m b e r 2003 2004 2005 2006 Time C ra s h F re q u e n c y

Figure 9. Rural Interstate Fatal Crash Frequency

The 70-MPH speed limit: speed adaptation, spillover and surrogate measures of safety